I bought my copy before I took a serious interest in earth science. When I first realised that AGW was a real threat, that book was the only one of which I knew where a serious scientist had expressed doubts about those dangers. Rereading it now, I see that what he wrote may explain why my sea ice prediction this year went so wrong.
He wrote that the dispute over global warming was between those who thought that feedbacks played a major role (e.g. Jim Hansen) and those who did not (e.g. Dick Lindzen.) To explain feedbacks he uses as an example a toy box that was popular in the early 1980's. He wrote
"The best example of feedback that I have ever seen was a black box on the side of which was a toggle switch and a sign: DO NOT TURN ON. Perhaps you could resist the temptation, but I couldn't. The result was, after some grinding and whirring from inside the box, that the lid opened, a hand emerged, turned off the switch, and then withdrew back into the box. Turning on the switch eventually led to a feedback which eventually turned it off. And so it is also with the atmosphere. One possibility, for example, might be that warming the oceans will result in more water vapour in the atmosphere and, as a consequence, more clouds. But this might lead to more solar radiation scattered to space, hence lower temperatures."
As an electrical engineer, I did not feel that this was a good example of feedback. There are two types of feedback - positive and negative. This box is a horrible example of both. The small mechanical operation of a switch is amplified to become a lid opening and the hand emerging, so showing positive feedback. But the switch being operated and the system returning back to its original state is a negative feedback. The two types of feedback have been combined in an unholy alliance.
I also found his atmospheric example flawed. If warming the oceans causes more clouds, and clouds cause cooling, then when that cooling occurred the clouds would disappear and the ocean surface would warm again. Therefore, if more atmospheric CO2 causes the sea surface to warm, and more clouds to form, then we cannot have a net cooling otherwise there would be less clouds than before and the planet would warm again.
However what is possible is that one summer we have the oceans warmed by clear skies, and the following summer these warm seas create more clouds cooling the oceans. The result would be a biennial oscillation in sea surface temperatures. Of course the cycle could take longer than two years, and then that describes ENSO, the El Nino Southern Oscillation. During an El Nino the Pacific Ocean is covered in cloud and the water which has accumulated in the Warm Pool loses its heat to the atmosphere.
So this is where I went wrong with My Arctic sea ice prediction. I was arguing that the positive feedback from ice-albedo effect must drive on the continuous reduction in the size of the sea ice extent. However, an unknown and invisible hand has emerged and switched off the melting process. In fact it is probably the one that Craig Bohren mentioned - clouds. The thinner ice in the Arctic is forming more leads and so releasing more water vapour into the air. This is forming clouds which is preventing the sun from having its full effect on the ice, so inhibiting the ice-albedo effect. But as I explained, this mechanism cannot reverse the sea ice retreat, only slow it.
If we finally emerge from the current solar minimum, will next year see another major sea ice retreat?

